System and method for WiFi-based indoor localization via unsupervised domain adaptation

Abstract

A location-aware electronic device is provided. The electronic device trains feature extraction layers, reconstruction layers, and classification layers. The training may be based on a reconstruction loss and/or a clustering loss. The electronic device processes a fingerprint to obtain an augmented fingerprint using randomization based on statistics of the fingerprint. The feature extraction layers provide feature data to both the reconstruction layers and the classification layers. The classification layers operate on the codes to obtain an estimated location label. An application processor operates on the estimated location label to provide a location-aware application result to a person.

What is claimed is: 
     
       1. A method of location determination performed by an electronic device including a transceiver and a memory storing an AI model, the method comprising:
 receiving, via the transceiver, a first fingerprint associated with a first person and a first environment, wherein the first person is an unregistered user; 
 obtaining, based on the first fingerprint, a first set of feature data; 
 determining, based on the first set of feature data, a reconstruction loss associated with the first person and the first fingerprint; 
 in case of the reconstruction loss being less than or equal to a threshold:
 classifying, based on the first set of feature data, the first fingerprint to obtain a first estimated location label of the first person, and 
 outputting, an indicator of the first estimated location label of the first person; 
 
 in case of the reconstruction loss being greater than the threshold, performing a registration of the first person. 
 
     
     
       2. The method of  claim 1 , wherein the registration includes:
 obtaining, an identifier of the first person, 
 obtaining, a ground truth location label of the first person, and 
 obtaining, via the transceiver, a second fingerprint associated with the ground truth location label; and 
 wherein the method further comprises providing an output based on the first estimated location label or the ground truth location label. 
 
     
     
       3. The method of  claim 1 , further comprising training weights, based on a third fingerprint associated with the first environment and with a second person who is a registered user, to obtain updated weights. 
     
     
       4. The method of  claim 3 , wherein the training weights comprises:
 receiving, the third fingerprint associated with the first environment and with the second person; 
 augmenting the third fingerprint to form a first augmented fingerprint; and 
 updating weights of the system based on the first augmented fingerprint. 
 
     
     
       5. The method of  claim 3 , wherein the determining a reconstruction loss comprises:
 receiving, via the transceiver, the first fingerprint; 
 determining, based on the updated weights, the first set of feature data; 
 determining, based on the first set of feature data, a first reconstructed fingerprint of the first person; and 
 determining, based on the first fingerprint and the first reconstructed fingerprint, the reconstruction loss of the first reconstructed fingerprint. 
 
     
     
       6. The method of  claim 2 , wherein the output includes the electronic device automatically shutting off based on recognizing that a room is empty. 
     
     
       7. The method of  claim 2 , wherein the output includes beam-forming, by the electronic device, sound to the first estimated location label of the first person. 
     
     
       8. The method of  claim 2 , wherein the output includes pan-tilting, by the electronic device, a display screen of the electronic device depending on the first estimated location label of the first person. 
     
     
       9. The method of  claim 2 , wherein the output includes updating an augmented reality presentation to the first person based on the first estimated location label of the first person. 
     
     
       10. The method of  claim 2 , wherein the output includes providing an advertisement to a personal electronic device of the first person based on the first estimated location label of the first person. 
     
     
       11. The method of  claim 1 , wherein the AI model includes a neural network based AI model, the method further comprises:
 jointly training, based on a reconstruction loss and a clustering loss, feature extraction layers, reconstruction layers, and classification layers of the neural network based AI model, wherein the reconstruction layers are configured to provide the reconstruction loss, and the classification layers are configured to provide the clustering loss; 
 processing a fingerprint to obtain an augmented fingerprint; 
 operating the feature extraction layers on the augmented fingerprint to produce codes; 
 classifying, with the classification layers, the codes, to obtain an estimated location label; and 
 providing an output based on the estimated location label. 
 
     
     
       12. The method of  claim 11 , wherein the output includes at least one of the electronic device automatically shutting off based on recognizing that a room is empty; beam-forming, by the electronic device, sound to the estimated location label; pan-tilting, by the electronic device, a display screen of the electronic device depending on the estimated location label; updating an augmented reality presentation to a person based on the estimated location label; and providing an advertisement to a personal electronic device of a person based on the estimated location label. 
     
     
       13. An electronic device comprising:
 one or more memories, wherein the one or more memories include instructions and store an AI model; 
 a transceiver; and 
 one or more processors, wherein the one or more processors are configured to execute the instructions to:
 receive, via the transceiver, a first fingerprint associated with a first person and a first environment, wherein the first person is an unregistered user, 
 obtain, based on the first fingerprint, a first set of feature data, 
 determine, based on the first set of feature data, a reconstruction loss associated with the first person and the first fingerprint, 
 in case of the reconstruction loss being less than or equal to a threshold:
 classify, based on the first set of feature data, the first fingerprint to obtain a first estimated location label of the first person, and 
 output, an indicator of the first estimated location label of the first person, 
 
 in case of the reconstruction loss being greater than the threshold, perform a registration of the first person. 
 
 
     
     
       14. The electronic device of  claim 13 , wherein the one or more processors are further configured to execute the instructions to:
 perform the registration by:
 obtaining, an identifier of the first person, 
 obtaining, a ground truth location label of the first person, and 
 obtaining a second fingerprint associated with the ground truth location label; and 
 
 provide an output based on the first estimated location label or the ground truth location label. 
 
     
     
       15. A non-transitory computer-readable storage medium storing instructions configured to cause a processor to:
 receive, via a transceiver, a first fingerprint associated with a first person and a first environment, wherein the first person is an unregistered user of an electronic device; 
 obtain, based on the first fingerprint, a first set of feature data; 
 determine, based on the first set of feature data, a reconstruction loss associated with the first person and the first fingerprint; 
 in case of the reconstruction loss being less than or equal to a threshold:
 classify, based on the first set of feature data, the first fingerprint to obtain a first estimated location label of the first person, and 
 output, an indicator of the first estimated location label of the first person; 
 
 in case of the reconstruction loss being greater than the threshold, perform a registration of the first person, by:
 obtaining, an identifier of the first person, 
 obtaining, a ground truth location label of the first person, and 
 obtaining a second fingerprint associated with the ground truth location label; and 
 
 provide an output based on the first estimated location label or the ground truth location label. 
 
     
     
       16. An apparatus comprising:
 an input configured to receive a first plurality of fingerprints and a second plurality of fingerprints, wherein the first plurality of fingerprints includes a first fingerprint, and the first fingerprint is associated with a first person at a first location, and wherein the second plurality of fingerprints includes unlabeled data and a second fingerprint; 
 a data augmenter comprising at least one first processor and at least one first memory, wherein the data augmenter is coupled to the input, and is configured to generate variational data based on the first plurality of fingerprints; 
 a feature extractor comprising at least one second processor and at least one second memory, coupled to the data augmenter, wherein the feature extractor is configured to provide first codes and second codes, the first codes are associated with the variational data, and the second codes are associated with the second plurality of fingerprints; 
 a reconstructor comprising at least one third processor and at least one third memory, coupled to the feature extractor, wherein the reconstructor is configured to produce a reconstructed first fingerprint and a reconstructed second fingerprint; 
 a classifier, comprising at least one fourth processor and at least one fourth memory, coupled to the feature extractor, wherein the classifier is configured to predict a second location of a second person, wherein a statistical performance measure of a prediction of the second location of the second person is optimized in a location label domain; and 
 a controller comprising at least one fifth processor and at least one fifth memory, wherein the controller is configured to:
 perform a first update of classification weight of the classifier, feature extraction weight of the feature extractor, and reconstruction weight of the reconstructor, based on a first difference between the reconstructed first fingerprint and the first fingerprint, and 
 perform a second update of the feature extraction weight and the reconstruction weight, based on a second difference between the reconstructed second fingerprint and the second fingerprint. 
 
 
     
     
       17. The apparatus of  claim 16 , wherein the second update is a joint update. 
     
     
       18. The apparatus of  claim 16 , wherein the second update is based on a clustering loss measure, wherein the clustering loss measure promotes avoidance of high-density regions when adjusting the classification weight. 
     
     
       19. The apparatus of  claim 16 , wherein the controller is further configured to perform a third update of classification weight, based on a difference between the reconstructed second fingerprint and the second fingerprint.